Results 1 ... 60 found in all logged channels for 'f:asc whitening'
(asciilifeform) asciilifeform: jonsykkel: you're probably thinking of this one
(asciilifeform) asciilifeform: whaack: note that bias in the usual sense aint the only possible rng failure. there's also, e.g., sequential correlation.
(asciilifeform) asciilifeform: raw_avocado: http://logs.nosuchlabs.com/log-search?q=f%3Aasc+whitening&chan=all
(asciilifeform) asciilifeform: punkman: his rng article also lulzy, where he shits out kilometre of sophistry to pretend that he aint whitening
(asciilifeform) asciilifeform: is why, for instance, there are 'over 9000' whitening pseudo-trngs on the market today, but the only nonwhitening, shielded trng had print run of 200 , half of which was sold at the bankruptcy hammer
(asciilifeform) asciilifeform: it's a lemon market, the homeopaths will always control the chump-change segment, while the crypto-ag types will control the golden-toilet segment. while actual rng that actually non-whiteningly worx, 'doesn't exist, never existed, what kinda terrorist are you to ask'
(asciilifeform) asciilifeform: genning with gpg, however, is problematic.
(asciilifeform) asciilifeform: and recall this.
(asciilifeform) asciilifeform: verisimilitude: be aware of known bugs .
(asciilifeform) asciilifeform: for all the surface simplicity of trng, it is surprisingly hard to get 100% right, esp. if you are unwilling to lie to yourself & others via 'whitenings'.
(asciilifeform) asciilifeform: shinohai: i've many interesting things lined up. incl. for phuctor. such as e.g. previously-unmolested classes of weak key .
(trilema) asciilifeform: why this whitening sadness.
(trilema) asciilifeform: for that matter, why not let user give the full 528bytes (132 x 32b) of key material, why this whitening is hardwired in.
(trilema) asciilifeform: i was about to add 'next will be usg.clone pseudo-FUCKGOATS' but remembered , they already released several , all of course with whitening and voltage booster pissing oscillation into power rail , and afaik all 2 (3? 4?) of'em even less popular than FG ( because why to buy whitened turdrng, intel ships for free on-die )
(trilema) asciilifeform: since the debian incident, enemy stepped up the 'NOBUS' crapola; no noar '32768 possible keys, total', instead things moar in the spirit of http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg
(trilema) asciilifeform: as for example fg discards whitening
(trilema) asciilifeform: ( will also point out, the lamps per se contribute ~0 entropy, arrangement is really ~same as hashwhitening output of camera static with the lens cap on )
(trilema) asciilifeform: ( and problematic in other ways -- not even shy about using 'whitening' for instance. )
(trilema) asciilifeform: Barbarossa_: see also http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg
(trilema) asciilifeform: 11:07 < kanzure> well i would prefer star shades, but cloud whitening seems to be popular.
(trilema) asciilifeform: iirc the d00d who found the koch whitening lulzgem used a proggy that worked quite like 'barium enema'
(trilema) asciilifeform: re whitening, it might be interesting to apply 'dragonfly fpga' to 'distinguish sha'd bitstream from nonsha'd'
(trilema) asciilifeform: 'whitening' and all the other sins, flow from it
(trilema) asciilifeform: phf: ... or where the author elaborately justifies 'whitening'
(trilema) asciilifeform: 'whitening'
(trilema) asciilifeform: which is 'whitening' if the starting material is 0000....000 , but doesn't hurt if trng.
(trilema) asciilifeform: doesn't advocate whitening, or any other obvious liquishit, which is a jawdropping rarity.
(trilema) asciilifeform: whitening == gluing petrol tank gauge to 'full'
(trilema) asciilifeform: if i want whitening, i can dd if=/dev/urandom .
(trilema) asciilifeform: see also http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg .
(trilema) asciilifeform: this came up in the rng whitening thread, but bears mentioning now and again
(trilema) asciilifeform: so i hand-cranked the sequence using mircea_popescu's method, to agglomerate, forever, a record of the path through the tree. and he is right that this will keep the toposort from detecting a loop. but this is rather like whitening an rng, it confuses only the machine, but not the reader, who will see that the ~output~ resulting from following the path is identical.
(trilema) asciilifeform: 'Data coming from true random number generators is never 100% random. I am aware of no exceptions. Whitening is required in all cases before the data is suitable for use in cryptography.' << from first link
(trilema) asciilifeform: $s whitening
(trilema) asciilifeform: 'whitening.'
(trilema) asciilifeform: BingoBoingo: http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg/#comment-67939 << answr.
(trilema) asciilifeform: dafuq is 'fixed' whitening.
(trilema) asciilifeform: whitening is in a class of similar items, childhood fears coupled to faux techno-fixes, deserves own separate discussion one day imho.
(trilema) asciilifeform: the fundamental psychiatric reason why folks like whitening, is, i suspect, a demon that wakes them up at night:
(trilema) asciilifeform: whitening sux.
(trilema) asciilifeform: ;;later tell mircea_popescu http://www.loper-os.org/pub/20166313-pg1_2.png and http://www.loper-os.org/pub/20166313-pg2_2.png (depdfization of dorre and klebanov's report on gpg whitening idiocy)
(trilema) asciilifeform: dump is post-whitening.
(trilema) asciilifeform: ;;later tell mircea_popescu http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg/#comment-67543
(trilema) asciilifeform: during my audit of the rng routine, i barfed at the whitening and stopped reading.
(trilema) asciilifeform: http://people.item.ntnu.no/~danilog/Hash/Non-random-behaviour-narrow-pipe-designs-03.pdf << interesting re 'whitening' etc. (tldr: no known hash behaves like 'random oracle', demonstrably)
(trilema) asciilifeform: it is entirely like whitening.
(trilema) asciilifeform: e.g., 1 and 2 imply that there are sufficient bits per entropy without having to resort to whitenings or distillation
(trilema) asciilifeform: (and 'rng with whitening' which IS a prng)
(trilema) asciilifeform: this is quite like the 'rng whitening'.
(trilema) ascii_field: he asked 'i get 7.9999999 bts/byte from urandom, why not from yours' and we had a chat about the idiocy of whitening
(trilema) asciilifeform: there is the basic truth, and there is, on top of it, a massive 'field' of exercises in self-delusion (e.g., 'whitening.')
(trilema) asciilifeform: jurov: and yes, they are typically 'best practices' retarded - e.g., 'whitening' etc
(trilema) asciilifeform: 'whitening' an rng is mostly fooling yourself.
(trilema) asciilifeform: !s whitening
(trilema) asciilifeform: 'conditioning' normally called 'whitening'
(trilema) asciilifeform: (what rng? ring oscillator jitter, with sha1 whitening.)
(trilema) asciilifeform: 'any periodic noise that makes it through simply reduces the bits/byte number of the entropy and gets smeared out by subsequent whitening' << is he serious? gotta be trolling
(trilema) asciilifeform: Rassah: ask him if whitening is used anywhere.
(trilema) asciilifeform: Rassah: why are you using whitening (hashing) in rng? and 'because everyone does' is not acceptable answer.
(trilema) asciilifeform: Rassah, mircea_popescu: then one last question, for if/when he does. value of adjacent sram cells is correlated. what, if anything, was done about this? (my guess - nothing but the usual 'whitening'.)